Background
Prostate cancer is the most frequently diagnosed cancer in the world. Most prostate cancers are initially dependent on androgens for growth, and patients with prostate cancer receive hormonal therapy. Androgen deprivation by medical or surgical castration contributes significantly to disease control during early stages of prostate cancer; however, the effect is usually palliative, and a majority of prostate cancers eventually progress to a hormone-refractory phenotype against which current treatments are relatively ineffective [
1]. The progression of prostate cancer from the androgen-dependent to androgen-independent state is the main obstacle in improving the survival and quality of life in patients with advanced prostate cancer. Therefore, much attention has been focused on the evolution from androgen-dependent to androgen-independent prostate cancer, and the establishment of novel therapeutic strategies against hormone-refractory prostate cancer (HRPC) is desirable.
In the past, molecular mechanisms for the progression to the hormone-refractory state have been proposed based on experimental evidence. The androgen receptor (AR)-dependent mechanisms include (a) androgen-independent activation of AR [
2,
3], (b) AR overexpression or mutations, which could allow AR to respond to lower levels of androgens [
4] or be directly activated by other ligands [
5], (c) increased expression of steroidogenic enzymes [
6], and (d) indirect activation of AR by cell-surface receptors such as HER2 [
7], the interleukin-6 receptor [
8] and G-protein-coupled receptors [
9]. The AR-independent mechanisms include (e) mutations of tumor suppressor genes [
10], (f) expression of various oncogenes affecting cell growth and death [
11], (g) enhanced angiogenesis [
12], (h) bypassing the AR pathway [
13], and (i) prostate cancer stem cell regeneration [
14].
Recently, Lyons
et al. reported a novel ligand-independent AR activation through Rho guanosine triphosphatase (GTPase) signaling in prostate cancer
in vivo and
in vitro [
15]. The levels of Vav3, a Rho GTPase guanine nucleotide exchange factor (GEF), are elevated in human prostate cancer specimens, and they increase during the progression of prostate cancer to androgen independence by enhancement of AR transcriptional activity [
16,
17]. The Vav gene (Vav1) was first identified in hematopoietic cells with oncogenic activity. Since the discovery of the Vav oncogene, new family members have been identified in mammalian cells (Vav2 and Vav3) [
18‐
20]. The biochemical functions of Vav family proteins have been extensively investigated. Vav1 expression is restricted to hematopoietic cells, and it is involved in the formation of the immune synapse. Vav2 and Vav3 are more ubiquitously expressed [
20]. Vav proteins contain the Dbl homology domain, which confers GEF activity, as well as protein interaction domains that allow them to function in pathways regulating actin cytoskeleton organization [
21]. In particular, their GEF activity is the most important function among them. Vav3, a signal transducer of receptor protein tyrosine kinase, is involved in various cellular signaling processes including cell morphology modulation and cell transformation with oncogenic activity [
22]. In the current study, Vav3 was demonstrated to bind to phosphatidylinositol 3-kinase (PI3K), leading to PI3K activation with cell transformation activity [
23].
In a previous report,
Dong et al. found that Vav3 enhances AR activity partially through PI3K/Akt signaling and stimulates androgen-independent growth in prostate cancer [
17]. We further revealed that tumor cell hypoxia induced Vav3 overexpression with androgen-independent growth and malignant behavior in LNCaP cells [
24,
25]. Therefore, we hypothesized that Vav3 has an important role in regulating the growth and survival of prostate cancer cells under hypoxic conditions and that it is a novel therapeutic target for the treatment of HRPC. In recent years, taxane-based chemotherapy has contributed to improvements in treatment outcomes in prostate cancer, and docetaxel has become a standard chemotherapeutic agent for treating HRPC; however, docetaxel does not exhibit sufficient activity when administered as a single agent [
26‐
28]. However, when docetaxel is used in combination with other therapeutic modalities, this therapeutic strategy may provide meaningful improvements in the management of HRPC. In this study, we report studies assessing
in vitro and
in vivo combinations of docetaxel with small interfering RNA (siRNA) for Vav3.
To the best of our knowledge, we have reported for the first time that interrupting the Vav3 signaling pathway using siRNA induces apoptosis and enhances docetaxel sensitivity through the inhibition of PI3K/Akt, extracellular signal-regulate kinase (ERK), and AR signaling axis in human prostate cancer.
Discussion
Docetaxel is a microtubule-targeting drug currently used as a standard first-line chemotherapeutic agent for the management of HRPC that has contributed to improved survival and quality of life in patients with advanced prostate cancer; however, its effectiveness is limited by intolerance and the development of docetaxel-refractory prostate cancer [
26‐
28]. It is therefore reasonable to expect further improvements in treatment outcomes when docetaxel is combined with other therapeutic modalities active against prostate cancer. Because the Vav3 oncogene is overexpressed in androgen-independent prostate cancer, in which it regulates cell growth [
16,
17], verifying whether Vav3 is a signaling molecule appears beneficial for establishing a new therapeutic target for treating HRPC in combination with docetaxel. We first tested the anti-cancer potential of interrupting the Vav3 signaling pathway using siRNA followed by investigating the impact of si-Vav3 in combination with docetaxel. The goals of this study were to (1) explore Vav3 as a novel therapeutic target for human prostate cancer, (2) define the biological effects of si-Vav3 when combined with docetaxel in human prostate cancer cells in culture and experimental animal models, and (3) characterize the downstream signaling pathways of Vav3 in human prostate cancer cells. This approach allowed us to advance our understanding of the possible importance of Vav3 as an efficacious therapeutic modality for prostate cancer beyond its commonly described associations with cell morphology and transformation.
In the present study, we made certain observations. (1) Vav3 was overexpressed in LNCaP cells cultured under chronic hypoxia (LNCaPH) characterizing androgen independence [
25]. (2) Vav3 activated pro-survival signaling pathways, including the activation of PI3K/Akt and ERK, which caused downstream Bad and AR phosphorylation in LNCaPH cells. (3) Downregulation of Vav3 signaling pathways by siRNA in combination with docetaxel significantly inhibited LNCaPH cell growth through the induction of apoptosis
in vitro and in mouse xenografts
in vivo. (4) si-Vav3 inhibited the phosphorylation of Akt and ERK, resulting in the inhibition of Bad and AR phosphorylation. (5) Docetaxel also inhibited the phosphorylation of Akt and ERK but activated JNK, resulting in increased Bcl-2 phosphorylation, and decreased Bad phosphorylation. To the best of our knowledge, this is the first report to show that siRNA knockdown of Vav3 can be combined with docetaxel against prostate cancer to yield increased sensitivity
in vitro and
in vivo.
Recent studies have suggested controversies in the roles of hypoxic tumor microenvironment in prostate cancer. Dihydrotestosterone increased hypoxia response element (HRE)-mediated transcriptional activity in prostate cancer, and androgen is involved in the response to hypoxia through hypoxia-inducible factor-1α [
29]. In addition, castration therapy was reported to decrease the synthesis of vascular growth factors, such as VEGF and angiopoietins, and upregulate hypoxia, leading to apoptosis in prostate cancer [
30]. Therefore, androgen deprivation therapy, which induces apoptosis by degenerating the vascular support system of the tumor, is reasonable for androgen-dependent prostate cancer. In contrast, tumor hypoxia is progressively associated with increased AR activity, reduced oxidative defense, genomic instability, and apoptosis resistance, and it may be associated with the transition to androgen independence in prostate cancer [
31,
32]. Suzuki
et al. reported that prostate cancer progresses in hypoxic conditions and transforms to the androgen independent state by suppressing the androgen response [
33]. Moreover, Butterworth
et al. also demonstrated that hypoxia could select for androgen-independent prostate cancer cells with more malignant behaviors including invasion and metastasis [
34]. In other words, hypoxia may select androgen-independent prostate cancer with a more malignant phenotype. We also previously reported that chronic hypoxia markedly potentiated androgen-independent growth and malignant behavior in LNCaP cells [
25]. Hence, it appears important to overcome the hypoxia-induced malignant potential reflecting the androgen-independent state in prostate cancer.
Vav3 has been identified as a Ros receptor protein tyrosine kinase-interacting protein functioning as a signaling molecule downstream of Ros [
23]. Vav3 also plays a role in epidermal growth factor receptor-, insulin receptor-, and insulin-like growth factor-mediated signaling pathways [
23]. Lyons
et al. reported that Vav3 expression is elevated in prostate cancer specimens and is coupled to growth factor receptor pathways that are upregulated during the progression of androgen-dependent prostate cancer cells to the androgen-independent state [
15]. Because Vav3 expression in LNCaP cells was also increased after long-term androgen deprivation, the possibility that Vav3 expression plays a role in the acquisition of androgen independence was suggested by these observations. Our previous study revealed that androgen-dependent LNCaP cells could acquire androgen independence through Vav3 overexpression when cultured under chronic hypoxia [
24,
25]. That is, prostate cancer under chronic hypoxia may reflect the androgen independent state with Vav3 overexpression.
We hypothesized that Vav3 may be a key therapeutic target molecule in the regulation of prostate cancer growth and survival under chronic hypoxia. To test this hypothesis, we examined the effects of Vav3 depletion by siRNA on cell growth and downstream cell signaling pathways in LNCaPH cells. We demonstrated that si-Vav3 alone inhibited LNCaPH cell growth and induced apoptosis
in vitro and in mouse xenografts
in vivo. These results are consistent with previous observations reported by Dong
et al., in which Vav3 depletion by siRNA inhibited growth in both androgen-dependent and androgen independent prostate cancer [
17]. However, the effect of si-Vav3 was weak and this study was designed to determine the combinatorial effects of docetaxel on cancer cell growth and apoptosis. In this study, we noted that the growth-inhibitory effect of si-Vav3 on LNCaPH cells occurred through a decrease in phosphorylated Akt and ERK, leading to the induction of apoptosis (Figures
2 and
3). Accompanying this apoptotic induction, we observed that si-Vav3 could induce caspase-9 activation (Figure
3) but not casapase-8 activation (data not shown). Taken together, these results suggest that si-Vav3-induced apoptosis mainly depends on mitochondrial pathways rather than death receptor-mediated pathways. In addition, combination treatment significantly decreased the phosphorylation of Akt and ERK and increased the phosphorylation of JNK. This indicates that combined si-Vav3 and docetaxel treatment increased apoptosis by modulating Akt, ERK, and JNK phosphorylation.
si-Vav3 induced the apoptotic activity of the pro-apoptotic member of the Bcl-2 family protein Bad, which is a common target of Akt and ERK [
35,
36]. We noted that inhibition of survival kinase cascades targeting Bad, including the inhibition of Akt and ERK phosphorylation by si-Vav3, resulted in apoptosis (Figures
2 and
3). In addition, si-Vav3 simultaneously inhibited the androgen signaling mediated by AR, a ligand-activated transcription factor and survival factor. It has been amply documented that AR, PI3K/Akt, and ERK pathways are important features contributing to uncontrolled prostate cancer cell growth and survival. In addition, increased AR activity is caused by crosstalk between AR and multiple intracellular signaling cascades, particularly PI3K/Akt and ERK pathways [
17,
37]. Consistent with a previous report, Vav3 downregulation inhibited AR phosphorylation through its convergent signaling network of PI3K/Akt and ERK in LNCaPH cells. Because PI3K/Akt and ERK pathways can regulate prostate cancer survival through the AR pathway, we investigated whether inhibiting PI3K/Akt and ERK pathways by kinase-specific inhibitors could affect AR phosphorylation. We found that treatment with LY294002 or U0126 decreased the phosphorylation of AR with a concomitant reduction of Akt and ERK phosphorylation in both LNCaP and LNCaPH cells (Figure
4). Thus, treatment with LY294002 increased apoptosis, whereas the effect of U0126 on cell apoptosis was attenuated compared with that of LY294002 because of the low level of basal ERK activity. Interestingly, the effects of LY294002 on apoptosis was stronger in LNCaPH cells than in LNCaP cells because of the high basal Akt phosphorylation level. Collectively, these results indicate that the PI3K/Akt signaling pathway plays a crucial role in LNCaPH cell growth, although cancer cell growth regulated by Vav3, at least in part, originated from activated ERK signaling. Our observations support the view that the PI3K/Akt pathway, which is activated by Vav3, is mainly involved in AR activity in prostate cancer development and progression [
17,
38].
In this study, although docetaxel also induced LNCaPH cell apoptosis by the inhibition of Bad phosphorylation including the inhibition of Akt and ERK phosphorylation, the level of AR phosphorylation was unaffected by docetaxel. It has been reported that docetaxel can induce apoptosis by PI3K/Akt inhibition in prostate cancer [
39]. Similarly, our findings suggested that Bad phosphorylation is mainly regulated by the PI3K/Akt pathway because basal ERK activity is very low in LNCaPH cells, although ERK phosphorylation is inhibited by docetaxel.
JNK, also called SAPK, is involved in development, morphogenesis, cell differentiation, and cell death in response to various environmental stresses including mitogen growth factors, inflammatory cytokines, oxidative stress, and diverse extracellular stimuli including cytotoxic drugs. Depending on the stimulus and cell type, JNK has essential roles in modulating the function of the mitochondrial pro- and anti-apoptotic proteins, such as Bcl-2 and Bad [
40,
41]. A previous study demonstrated that the duration and intensity of JNK activation are associated with apoptotic cell death and that Bad dephosphorylation is accompanied by increases in JNK phosphorylation and activity [
42]. Moreover, JNK activation leads to inactivation of the survival functions of Bcl-2 through Bcl-2 phosphorylation [
43]. In this study, Bad dephosphorylation and Bcl-2 phosphorylation with an elevation of JNK phosphorylation were induced by docetaxel, suggesting that JNK positively regulates apoptosis induction through Bad dephosphorylation and Bcl-2 phosphorylation.
Methods
Cell culture and hypoxia induction
LNCaP human prostate cancer cells were maintained in RPMI 1640 medium supplemented with 10% heat-inactivated fetal bovine serum (FBS) (Bio-Whittaker, Walkersville, MD, USA), 50 IU/ml penicillin, and 50 μg/ml streptomycin (GibcoBRL, Grand Island, NY, USA) and cultured at 37°C in a humidified atmosphere of 5% CO2. To establish chronic hypoxia-conditioned LNCaP (LNCaPH) cells, LNCaP cells were cultured under hypoxia (1% O2) for 6 months. The experiments using LNCaPH cells were performed under hypoxic conditions. KPK13 human renal cell carcinoma cells were cultured in minimum essential medium (MEM) (GibcoBRL) supplemented with 10% FBS, with 50 IU/ml penicillin, and 50 μg/ml streptomycin in 5% CO2 at 37°C.
Immunocytochemistry
Cultured cells were washed with PBS, fixed in methanol for 20 min, and incubated in 10% goat normal serum (Nichirei, Tokyo, Japan) for 10 min at 37°C. Cells were incubated in a primary antibody against Vav3 (1:50 dilution; Epitomics, Burlingame, CA, USA) at room temperature in PBS with 1% BSA for 60 min. After incubation with the primary antibody, the secondary antibody fluorescein dye (Alexa Fluor 488)-conjugated goat anti-rabbit IgG (1:40 dilution) (Invitrogen, Carlsbad, CA, USA) was added in PBS with 1% BSA for 30 min. Cells were visualized using confocal laser microscopy followed by nuclear staining with 1 μg/ml 2,4-diamidino-2-phenylindole dihydrochloride n-hydrate (Dojindo, Kumamoto, Japan).
Transient transfection of Vav3 siRNA
Cells were transiently transfected with a Vav3 siRNA duplex (si-Vav3; final concentration of 60 nmol/l) or a control siRNA (a random scrambled sequence: si-Scr; final concentration of 60 nmol/l) using Lipofectamine RNAiMAX (Invitrogen) according to the manufacturer’s instructions. The sequence of the siRNA against Vav3 synthesized by Invitrogen is 5′- AUAUUCUCCUGACUCUUUGGUCCUG(dT)(dT)-3′. Following transfection, cells were subjected to growth inhibition, live/death, flow cytometric, and immunoblot analyses.
Growth inhibition assay
Cell viability was determined using a cell proliferation assay. In brief, exponentially growing cells were seeded in 6-well plates at 1 × 105 cells/well. After overnight culture, the culture medium was changed to fresh standard medium containing 5 nM docetaxel (Sigma, St. Louis, MO, USA) for 0–72 h or various concentrations of docetaxel (0, 1, 2.5, and 5 nM) for 72 h in the presence or absence of si-Vav3. After treatment, the cell number was counted with a hemocytometer.
Live/death analysis
Cells were treated with si-Vav3, 5 nM docetaxel, or si-Vav3 plus 5 nM docetaxel for 48 h. Live and dead cells were detected using the Live/Death Viability/Cytotoxicity assay kit (Molecular Probes, Eugene, OR) for which fluorescence was observed and pictures were taken at 4× magnification. The data from three independent experiments were expressed as a mean percentage.
Flow cytometric and DNA fragmentation analyses
For cell cycle analysis, flow cytometric analysis of propidium iodide-stained nuclei was performed. In brief, cells treated with si-Vav3, 5 nM docetaxel, or si-Vav3 plus 5 nM docetaxel for 0–72 h or various concentrations of docetaxel (0, 1, 2.5, and 5 nM) for 72 h in the presence or absence of si-Vav3 were plated at a density of 5 × 105 cells in a 60-mm dish overnight. The cells were collected by trypsinization and fixed with 70% ethanol. The fixed cells were incubated with 100 μg/ml RNase A (Sigma) for 30 min and stained with 25 μg/ml propidium iodide (Chemicon, Temecula, CA, USA) for 30 min. Cell cycle distribution was analyzed with a FACScan flow cytometer and CellQuest software (Becton Dickinson Labware, Lincoln Park, NJ, USA). The data from three independent experiments were expressed as a mean percentage. The apoptotic response was also measured by the Cell Death Detection ELISAPLUS photometric enzyme immunoassay for the quantitative determination of cytoplasmic histone-associated DNA fragments (Roche, Indianapolis, IN, USA). In brief, the cytoplasmic fractions of the untreated control cells and cells treated with si-Scr, si-Vav3, 5 nM docetaxel, or si-Vav3 in combination with 5 nM docetaxel were transferred to a streptavidin-coated plate and incubated for 2 h at room temperature with a mixture of peroxidase-conjugated anti-DNA and biotin-labeled antihistone. The plate was washed thoroughly and incubated with 2, 2′-azino-di[3-ethylbenzthiazolin-sulfonate]. The absorbance was measured at 405 nm with a reference wavelength of 492 nm.
Cytotoxicity assay
To determine the involvement of PI3K/Akt, ERK, and c-jun N-terminal kinase (JNK) pathways in cell apoptosis, cells were treated with LY294002 (Calbiochem, Darmstadt, Germany), U0126 (Calbiochem), or SP600125 (Sigma), respectively, for 48 h. Control cells were cultured in the presence of an equivalent amount of DMSO as a vehicle.
Immunoblot analysis
Protein was extracted from cell pellets with a lysis buffer (50 mM Tris, pH 8.0, 150 mM NaCl, 0.02% NaN3, 0.1% SDS, 1% NP-40, 0.5% sodium deoxycholate, and 1 mM PMSF) in the presence of a protease inhibitor cocktail (Roche Applied Science, Indianapolis, IN, USA). Samples containing equal amounts of protein (30 μg) were electrophoresed on 8–16% Tris-glycine gels (Invitrogen) and transferred to nitrocellulose membranes. After blocking with T-TBS containing 5% nonfat milk powder, the membranes were incubated with mouse monoclonal antibody against phospho-Akt (Ser 473) (1:1000 dilution; Cell Signaling, Danvers, MA, USA), phospho-ERK (Thr 202/Tyr 204) (1:1000 dilution; Cell Signaling), phospho-stress−activated protein kinase (SAPK)/JNK (Thr 183/Tyr 185) (1:1000 dilution; Cell Signaling), and Bcl-2 (1:1000 dilution; Santa Cruz Biotechnology, Santa Cruz, CA, USA), or rabbit polyclonal antibodies against Vav3 (1:1000 dilution; Cell Signaling), Akt (1:1000 dilution; Cell Signaling), ERK (1:1000 dilution; Santa Cruz Biotechnology), phospho-Bcl-2 (Ser 70) (1:1000 dilution; Cell Signaling), Bad (1:1000 dilution; Cell Signaling), phospho-Bad (Ser 136) (1:1000 dilution; Cell Signaling), ERK (1:1000 dilution; Cell Signaling), SAPK/JNK (1:1000 dilution; Cell Signaling), AR (1:1000 dilution; Millipore, Temecula, CA), phospho-AR (Ser 81) (1:500 dilution; Millipore), caspase-3 (1:1000 dilution; Cell Signaling), caspase-9 (1:500 dilution; Cell Signaling), or poly(ADP-ribose) polymerase (PARP) (1:1000 dilution; Cell Signaling) at 4°C overnight. After washing with T-TBS, the membranes were incubated with corresponding secondary antibodies, which were conjugated with horseradish peroxidase (Santa Cruz Biotechnology). The blots were stripped and reprobed with anti-β-tubulin antibody (1:5000 dilution; Sigma). Immunoreactive bands were visualized with ECL plus (Amersham Pharmacia Biotech, Little Chalfont, UK) and quantified by scanning densitometry using NIH Image software (version 1.55).
Atelocollagen is a type I collagen of calf dermis that is highly purified by pepsin treatment (Koken Co., Ltd, Tokyo, Japan). The siRNA and atelocollagen complexes were prepared as follows. An equal volume of atelocollagen (in PBS at pH 7.4) and siRNA solution was combined and mixed by rotation at 4°C for 20 min. The final concentration of atelocollagen in vivo was 0.5%.
In vivo animal experiment
Four-week-old male athymic nude mice (20–24 g; BALB/c nu/nu mice) were housed in accordance with and approved by the Institutional Animal Care and Use Committee of Oita University. For subcutaneous injection, LNCaPH cells were trypsinized, and single-cell suspensions (2 × 106 cells) were mixed 1:1 with Matrigel and then injected into both flanks. To determine the optimal concentration of the siRNA/atelocollagen complex, dose–response tests including si-Scr as a vehicle control and si-Vav3 were performed. Three weeks after the injection of mice with LNCaPH cells, when the tumor volume reached 100 mm3, the mice were randomly divided into seven treatment groups (non-treatment; 0.5 μg, 2.5 μg, and 10 μg of si-Scr/tumor; 0.5 μg, 2.5 μg, and 10 μg of si-Vav3/tumor), each consisting of four mice. The siRNA/atelocollagen complex was injected directly into the tumors once a week for 7 consecutive weeks. Tumor size was quantified by measuring in two dimensions with calipers, and tumor volume was calculated every 7 days according to the equation (l × w2)/2, where l = length and w = width. The mice were monitored daily for changes in weight and other signs of acute toxicity. After optimizing the concentration of the siRNA/atelocollagen complex (2.5 μg siRNA/50 μl/tumor), the effects of combination therapy with docetaxel was assessed. Tumor cell-bearing mice were randomly divided into four treatment groups (si-Scr, si-Vav3, docetaxel, and docetaxel + si-Vav3), each consisting of four mice. An intratumoral injection of the siRNA/atelocollagen complex (si-Vav3 or si-Scr) was performed with or without docetaxel 10 mg/kg i.p. once weekly for 7 consecutive weeks. When treatments were completed, animals were sacrificed, and tumor tissues were harvested for immunoblot analysis, H&E staining and immunohistochemistry.
Immunohistochemical protocol
After sacrifice, s.c. tumor tissues were fixed with 10% buffered formalin and embedded in paraffin. The formalin-fixed, paraffin-embedded tissues were cut to 4-μm sections and deparaffinized in xylene followed by treatment with a graded series of ethanol and rehydration in PBS. The sections were incubated in 0.3% H2O2 for 10 min to inactivate endogenous peroxidase, followed by washing in PBS. To block non-specific binding to sections and eliminate non-specific staining, 10% normal goat serum in PBS was applied and incubated for 10 min. The following primary antibodies were used: Vav3, Ki-67 (DAKO, Carpinteria, CA, USA), phospho-AR (pAR) (Abnova, Thaipei, Taiwan), and M30 CytoDeath (Roche Applied Science, Mannheim, Germany). They were diluted 50×, 1×, 100×, and 50×, respectively, with PBS containing 1% BSA. After washing with PBS, sections were incubated with secondary antibodies, which were conjugated with peroxidase-labeled amino acid polymer (DAKO). The immune complex was visualized using a 3,3′-diaminobenzidine peroxytrichloride substrate solution (DAKO). Slides were then counterstained with hematoxylin and mounted. The evaluation of Ki-67, pAR, and M30 CytoDeath staining was based on the proportion of positive-stained cells among a total of 1000 cells that were counted in five randomly selected areas.
Statistical analysis
Values were expressed as means ± SE. Statistical analysis was performed using Student’s t-test. The limit for statistical significance was set at P <0.05.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
TN, MY, FS and HM designed this study and performed all experiments with the help of KH, TI and RS. KM, MM and HM critically read the manuscript and gave good advice. TN wrote this manuscript. All authors read and approved final manuscript.